CarbonAI Raises $12 Million to Automate Commercial Real Estate Lending

CarbonAI’s $12 million funding round puts a focused bet behind automating commercial real estate lending, where document review and credit analysis still rely heavily on manual work. The New York company says its AI agents handle underwriting, due diligence, investment memo drafting and post-close loan surveillance. The system is designed to take a loan file through connected stages and link its conclusions to the documents that support them.

Cowboy Ventures led the round. Timber Grove Ventures, Bling Capital and Max Ventures also participated. CarbonAI launched its first product suite roughly six months ago and says it has since reached seven figures in contracted annual revenue. That early commercial traction is notable, though the company hasn’t shared customer names beyond a reference from Greystone co-president Mordecai Rosenberg, who pointed to the platform’s due diligence automation.

A lending workflow under time pressure

The scale of upcoming loan maturities helps explain why lenders may be receptive to tools that shorten analysis cycles. The Mortgage Bankers Association estimates that $875 billion in commercial mortgages will mature in 2026. That volume puts pressure on teams reviewing loans for refinancing or restructuring, particularly as interest rate pressures prompt more borrowers to seek changes to existing financing.

The work involved is broad: lenders need to assess property income, lease terms, borrower financials and supporting documents, then translate those details into a credit view. CRE lending teams spend an estimated $50 billion per year on underwriting and loan monitoring, according to the source article. Much of that effort still runs through spreadsheets and manual document checks. Software that reduces repetitive review could free staff to focus on exceptions and credit judgment, but its usefulness depends on whether the outputs are accurate enough to rely on.

CarbonAI says its agents extract and organize information from rent rolls, leases, financial statements and legal documents. They then produce underwriting models and flag potential risks, with each output linked to its source document. The company also offers ongoing monitoring after a loan closes, replacing the traditional quarterly manual review cycle with continuous surveillance.

That source linkage is central to the product’s pitch. A model or risk flag is more useful when a lender can trace it back to the document and data behind it, especially when teams need to review assumptions or explain a credit decision internally. CarbonAI’s chief executive and co-founder, Jeff Saul, described the aim as completing the work from end to end rather than requiring users to guide an assistant through every task: “Lenders don’t need another chat window that they have to coach through every step. They need agents that already know the job, do it end to end consistently, and show their work.”

Domain experience and competitive pressure

Saul and co-founder Josh Kaufman have worked directly in the field they’re targeting. Both started their careers underwriting CRE loans at Starwood Capital, Morgan Stanley and Guggenheim. They later built Nativ, a commercial real estate deal management platform that served more than 30 lenders across $300 billion in assets under management before its sale to Built, a construction finance software company.

That background may help the founders identify where the workflow breaks down and which tasks can be standardized. It also gives CarbonAI a specific starting point as financial services firms test AI in document-heavy processes. General-purpose language models can handle a range of tasks, but lenders need systems that fit structured workflows, keep track of supporting material and produce work that can be checked. CarbonAI is positioning its document-linked outputs around those needs.

The company is entering a field that includes established data and analytics vendors adding AI features, along with newer firms focused on either loan origination or post-close monitoring. CarbonAI’s stated distinction is its coverage of the full loan lifecycle in one system. Cowboy Ventures partner Caroline Duffy cited the founders’ operating experience as a key reason for backing the company. That experience may open doors with lenders, while the product’s ability to perform consistently across different loan files will determine how much work customers entrust to it.

What the early traction can show

Greystone’s reference offers a concrete customer signal, but the company hasn’t disclosed a broader list of clients or account-level results. Its customers are described as debt funds, mortgage REITs, agency multifamily lenders and publicly traded financial institutions. Those groups have different processes and loan portfolios, so adoption across them could test whether the product travels beyond a narrow set of use cases.

The next evidence to watch is customer retention, revenue growth and how far CarbonAI’s agents move into decisions that shape lending outcomes. As automation takes on more of the analytical workflow, lenders will need to understand where human review remains necessary and how the system handles incomplete or conflicting information. For CarbonAI, the $12 million round supplies capital to build on early demand; durable growth will depend on whether customers continue using the platform and expand its role across the life of their loans.